OWNER’S MANUALcrosstalk.arizona.edu/Manuals/Timewarp.pdfEdgar Winter, Paul Bley, Roger Powell, Jean-Michel Jarre, Mike Oldfi eld, Herbie Hancock, and many, many others. Its modular

OWNER’S MANUAL

Owner’s Manual

1Copyright © 2004, by Way Out Ware, Inc.. All rights reserved.

No part of this work may be reproduced or transmitted in any form or by any means, electronic or mechanical, from or to any form of media, without the prior written permission of Way Out Ware, Inc..

Requests for permission to reproduce any part of this work should be addressed to : Way Out Ware, Inc., attn: Copyright Adminstration, [email protected].

1Table of Contents

1 The ARP 2600, 1970 – 1981 and onward... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

2 System Requirements, Installation, Confi guration, Setup and Usage . . . . . . . . . . . . . . . . . . 5

In this chapter you will fi nd all of the platform-dependent information you need in order to install and operate your TimewARP 2600 software synthesizer.

3 The Craft of audio Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

This chapter is about the facts – physical, mathematical, and auditory – that make the TimewARP 2600, and the hardware that it emulates, possible. We have to spend a few minutes here distinguishing between physical signals, and the sounds that people hear in the presence of certain kinds of signals.

4 Modular Components of the TimewARP 2600 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

In this chapter you’ll fi nd detailed explanations of the TimewARP 2600’s features and functions.

5 Patching the TimewARP 2600. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

6 Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

Table of Alternate keyboard tunings compiled by Robert Rich

7 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

The ARP 2600, 1970 – 1981 and onward... 1

The ARP 2600 was the second product of ARP Instruments. It was released in 1970, and continued until the manufacturer ceased operations in 1981.

Its design combined modularity (for studio fl exibility, and for instructional use) and modularity (for studio fl exibility, and for instructional use) and modularity integration (for realtime performance). Functionally, the ARP 2600 was completely modular: any signal output could be routed to any signal input, with a patch cord. Operationally, the ARP 2600 was integrated, using internally-wired default signal paths that made it possible to create a wide range of keyboard patches by simply opening up slide attenuators, as though sitting in front of a mixing console.

The ARP 2600 earned an early reputation for stability, and for fl exibility exceeding that of its competitor the Minimoog. Used 2600’s in good condition command premium prices on eBay today and businesses around the country can make a living reconditioning, rebuilding, and customizing 30-year-old units.

Among rock musicians, the ARP 2600 was used by Stevie Wonder, Pete Townsend, Joe Zawinul, Edgar Winter, Paul Bley, Roger Powell, Jean-Michel Jarre, Mike Oldfi eld, Herbie Hancock, and many, many others.

Its modular design, and the popularity of its Owner’s Manual, made the ARP 2600 a widely used teaching instrument in many schools and music conservatories around the country and internationally.

We are proud to bring you this software emulation of the 2600. Have fun with it, learn from it, but above all, make noise with it.

2

CHAPTER 1 - The ARP 2600 CHAPTER 1 -

Foreword

Unfasten the seat belts of your mind. The TimewARP 2600 will be an astonishing, exhilarating, and enlightening experience.

Creating this manual has been an astonishing, exhilarating, and enlightening experience for me. How many are ever given the chance to revisit an earlier life, an earlier project, a project like the ARP 2600 Manual, decades later, and get it right? It’s time travel. I’m grateful to Way Out Ware for providing me that opportunity.

When, at Alan R. Pearlman’s invitation, I began work on the original 2600 manual in September of 1970, the 2600 itself barely existed. For the fi rst two months, I was writing “blind” - without a machine in front of me. My fi rst hands-on experience with a synthesizer had been only six months earlier (it was a Putney VCS3).

I fi nished the text in March of 1971, Margaret Friend created the graphics, and the Owner’s Manual for the ARP 2600 began what turned out to be a surprisingly long career. In spite of the many defects that my inexperience contributed – the gaps in coverage, and outright errors - it became quite popular. To this day, it still gets an occasional respectable mention in the analog-synthesis community.

When Way Out Ware’s Jim Heintz called, early in 2004, to tell me about the TimewARP 2600, a lot of time had passed. Regarding software synthesizers, I had grown weary and cynical. Analog-modeling software had been a decade-long disappointment; some products did interesting things but not the things that real analog modules do. Jim, however, had already encountered, and thought about, and solved, these problems. He owned a real 2600. He really aimed at getting it right and would not be satisfi ed with anything less. It was a pleasure, fi nally, to accept his invitation to do an Owner’s Manual.

It’s clear, now, that Way Out Ware has set a new standard for software-based audio synthesis. The behavior of the TimewARP 2600 software – both module-by-module and integrated into patches - is effectively indistinguishable from that of the analog hardware that it emulates.

Soaring and swooping through the free air of analog synthesis – a world of nothing but sliders and cords and continuously evolving patch confi guration - was a capstone course at the Boston School of Electronic Music in the 1970’s. That is the world that the TimewARP 2600, for a new generation of musicians in a new millennium (that means you), provides access to: it is the fi rst – and I believe only - software synthesizer to support real-time performance by sliders and patchcords alone.

So here it is: your new Owner’s Manual, for the new TimewARP 2600.

Unfasten the seat belts of your mind. How else can you hope to experience time travel? How else can you enjoy free fl ight?

Jim Michmerhuizen

Jim Michmerhuizen is the author of the original ARP 2600 Manual and Founder and Director of the Boston School of Electronic Music.

CHAPTER 1 -

3

CHAPTER 1 - The ARP 2600

How this Manual is Organized

This is not a textbook; it’s a survival manual.

Chapter 2 is about installing and confi guring the software so you can get up and running.

Chapter 3 is a brief introduction to the vocabulary and methods of classical analog synthesis, so that we can understand each other throughout the rest of the book.

Chapter 4 is a module-by-module reference, including the digital extensions made possible by the fact that the TimewARP 2600 is software – a piece of computer behavior – rather than a collection of electronic hardware.

Chapter 5 is not in this book, but is a collection of patches, with accompanying documentation, located in a separate document fi le called Patchman.pdf found on the distribution CD (or as a download from Patchman.pdf found on the distribution CD (or as a download from Patchman.pdfwww.wayoutware.com). The patches and commentary are keyed to the numbered chapters, sections, and subsections of this manual. Some of the patches form a tutorial sequence, and some illustrate vocabulary lessons and concepts.

How To Use This Manual

You’ll probably need to do a quick run through of Chapter 2 as you install the TimewARP 2600 and learn some of the basic setup operations. After that, you can pretty much mix and match, according to your experience:

If you’re new to audio synthesis, you’ll want to walk through the tutorial patch sequence inPatchman.pdf, referring back to Chapter 3 for concepts and Chapter 4 for detailed module specifi cations while you explore, and learn to control, the vast range and patch repertoire of the TimewARP 2600 software synthesizer.

If you already have some experience with audio synthesis, you might go directly to Chapter 4 for the detailed module-by-module specifi cations of the TimewARP 2600. If you know and love the original ARP 2600 itself, take particular note of section 4.1, where we describe what you can do with the TimewARP 2600 that you could not do with its analog ancestor. These digital extensions include patch load/save, additional VCO sine-wave outputs, dual-channel signal input, automatic Y-connections at all signal outputs, sixteen keyboard micro-tuning options, MIDI Beat Clock synchronization, and MIDI controller mapping (with subranges) for all the panel sliders.

System Requirements, Installation, Confi guration, Setup and Usage 2

In this chapter you will fi nd all of the platform-dependent information you need in order to install and operate your TimewARP 2600 software synthesizer.

The TimewARP 2600 provides an extended set of digitally-based features that no hardware-based analog synthesizer can offer. We describe the most important of these in section 2.2.

2.1 Digidesign Pro Tools

This release of the TimewARP 2600 is an RTAS plug-in and runs only under Digidesign Pro Tools 6.1 and later versions, on either Apple OSX or Windows XP. On either of these platforms, the requirements for the TimewARP 2600 are essentially those of Pro Tools itself.

2.1.1 System Requirements

These are the same as Pro Tools itself, including enough memory to satisfy the practical needs of the TimewARP 2600 software.

2.1.1.1 Disk Space and RAM

The executable fi le occupies 6.1MB on disc, and the RAM requirements when running are the same as the Pro Tools minimums.

2.1.1.2 System Clock

The TimewARP 2600 performs all signal generation and processing in real time. In complex polyphonic patches, this may put a considerable load on your CPU. You may have diffi culty with a clock speed less than 800MHz; higher clock rates will increase the number of polyphonic voices you can use, and the complexity of the patches available.

2.1.2 Installation

Installation on either platform is accomplished by very simple standardized installation sequences. You must, of course, have a current installation of Pro Tools on your target computer.

6

CHAPTER 2 - System Requirements, Installation, Confi guration, Setup and Usage CHAPTER 2 -

2.1.2.1 Apple OSX

Insert the distribution CD or download the .dmg installer fi le from www.wayoutware.com. Double click the .dmg installer to initiate installation.

If you agreed to the license, a TimewARP 2600 installer icon appears. Double-click it to initiate the installation process.

When fi nished, start Pro Tools and proceed to 2.1.3.1.

2.1.2.2 Windows XP

Insert the distribution CD or download the installer fi le from www.wayoutware.com. Run the .exe installer fi le.

If you agreed to the license, follow the installer instructions.

When fi nished, start Pro Tools and proceed to 2.1.3.1.

2.1.3 Setups

2.1.3.1 Real-time Instrument for Performance and Recording

The TimewARP 2600 can transform your computer into a real-time instrument for live performance or recording.

To set this up, create an audio track in Pro Tools. Select either mono or stereo in the track-creation window; this will determine, in turn, the channel options offered by the TimewARP 2600 for this track.

In the mix window display for this track, at the track-inserts block up at the top, click on the fi rst insert selector. Choose the TimewARP 2600 plug-in from the appropriate menu.insert selector. Choose the TimewARP 2600 plug-in from the appropriate menu.insert

Now create another track, but in the track creation dialog, choose MIDI track instead of audio track, and route the track output to the TimewARP 2600.

As input for the new track, select your MIDI keyboard device, and enable the track for input for the new track, select your MIDI keyboard device, and enable the track for inputrecording. Pro Tools will route all incoming MIDI events from your MIDI keyboard device to the TimewARP 2600 plug-in.

If your MIDI keyboard has any additional MIDI control devices (sliders, knobs, buttons, etc.), you can assign these to any TimewARP 2600 sliders, knobs and switches that you choose. For details, see section 2.2.

2.1.3.2 Processing Audio Signals

The TimewARP 2600 can be used to process audio signals in a similar manner to a reverb or other effects plug-in.

To set this up, create an audio track or select a track already recorded in Pro Tools. (If this is a new track, select either mono or stereo in the track-creation window; this will determine, in turn, the channel options offered by the TimewARP 2600 for this track.)

In the mix window display for this track, at the track-inserts block up at the top, click on the fi rst available insert selector. Choose the TimewARP 2600 plug-in from the appropriate menu.

CHAPTER 2 -

7

CHAPTER 2 - System Requirements, Installation, Confi guration, Setup and Usage

If the current track is mono, the TimewARP 2600 for this track will be confi gured for one channel of input, and you can select between mono and stereo output. If the current track is stereo, the TimewARP 2600 insert will automatically be confi gured for stereo throughout. Confi rm this, when the synthesizer panel comes up, by observing that the preamp module in the upper left corner has two channel outputs rather than just one.

If you are processing an existing track, press play on the transport control to hear the audio being processed by the TimewARP 2600 in real time. You need to select an appropriate TimewARP 2600 patch (such as those in the Voice or Guitar Effects categories in the Factory group) to successfully process audio this way.

If you are processing a live track, enable the track’s record mode, and then audio from your input will be fed through the TimewARP 2600. You need to select an appropriate TimewARP 2600 patch (such as those in the Voice or Guitar Effects categories in the Factorygroup) to successfully process audio this way.

2.1.3.3 MIDI Tracks

Running as a track plug-in, the TimewARP 2600 can process prerecorded MIDI tracks.

To do this, set up two tracks just as you did above in section 2.1.3.1: an audio track that is running the TimewARP 2600 as a track insert, and a MIDI track whose output is routed to the TimewARP 2600.

When you play this track (use the Pro Tools transport window to play, rewind, pause the MIDI sequence, etc.), the MIDI events stored in the track sequence will be routed to the TimewARP 2600. MIDI note-events will become key depressions at the virtual keyboard; MIDI pitch-bend will turn the virtual keyboard pitch-bend knob; and MIDI controllers that are mapped to TimewARP 2600 sliders, knobs or switches (see section 2.2) will move those controls.

2.1.4 Selecting Patches

The TimewARP 2600 gives you a three-level hierarchy for storing and organizing your patches. All Patches are sorted into various Categories, which are in turn sorted into major Groups. Each of the three patch selection buttons generates a drop-down list associated with one layer in this hierarchy.

Groups, Categories, and Patches can also be selected by keyboard shortcuts. The up/down arrow keys on the computer keyboard select Patches, the left/right arrow keys move between Categories, and using the control key with the control key with the control key left/right arrow keys moves between Groups.

2.1.5 Making Patch Connections

Use the mouse to connect any output to any input jack. Position the cursor at any signal output, click and hold down the mouse button, and drag the patch cord to any signal input. To remove a patch cord, drag either end away from its signal jack.

The TimewARP 2600 will not connect two outputs or two inputs together. If you drag from one jack to another but no patch cord appears, it may be that both jacks are inputs or both are outputs.

8

CHAPTER 2 - System Requirements, Installation, Confi guration, Setup and Usage CHAPTER 2 -

The TimewARP 2600 will allow Y-connections, distributing a single output signal to multiple destinations by holding the control key down while clicking an output that already has a cable connected to it. A second cable will appear attached to the cursor, which can then be plugged into another input.

2.1.6 Using the Sliders and Knobs

To adjust a slider or knob, drag it with the mouse. For increased resolution, hold down the command (Apple) or command (Apple) or command control (Windows) key during the drag operation.control (Windows) key during the drag operation.control

If you pause for a moment over a slider, a display will pop up with the numerical value and name of the parameter it controls.

If you have MIDI controller hardware, you can easily connect it to the sliders, knobs or switches; see section 2.2.2.

2.1.7 Using the Virtual Keyboard Display

The graphic TimewARP 2600 keyboard display responds directly to mouse events; click on any key to create a MIDI keydown signal. This is useful when you are creating and tuning a new patch, if you don’t have a hardware MIDI keyboard handy.

The virtual keyboard also displays – as key depressions - incoming MIDI note-events. Use this to monitor your external MIDI keyboard connection and activity.

2.1.8 Polyphonic Operation

The TimewARP 2600 can respond to as many as eight simultaneous keydown events. Use the Voices dropdown (see section 4.1.1.2) to set the number of voices.

2.1.9 Saving and Loading Patches

You may save patches using the Save or Save As buttons or load patches with the Patch Manager (see section 4.1 below). There is no limit to the number of patches you can create and save.To enable the Save button, the Lock button (padlock icon) must be in the unlocked mode.

You may wish to create a bank of “template” patches, generic versions of often-used confi gurations.

The TimewARP 2600 responds to MIDI bank-select and patch-select commands.

2.2 MIDI Communication and Control

2.2.1 Slider, Knob and Switch Control

You may control any slider, knob or switch on the TimewARP 2600 plug-in with any MIDI controller. If you have a hardware control surface, or if your MIDI keyboard has slider or knob control devices, you can use them to control any combination of sliders, knobs or switches in a patch.

Any connections you set up are global to the TimewARP 2600 – you can store them independently of any specifi c patch (see section 4.1.2.1 below).

To connect a slider, knob, or switch to an external MIDI controller, hold down the control

CHAPTER 2 -

9

CHAPTER 2 - System Requirements, Installation, Confi guration, Setup and Usage

(Mac) or shift (Windows) key and click on the slider, knob or switch image. The responding shift (Windows) key and click on the slider, knob or switch image. The responding shiftdialog box offers you both global and patch-specifi c connections.

2.2.1.1 Global MIDI Map Settings

Select the controller number, either by directly controller number, either by directly controller numbertyping it in, or by twiddling the physical control on the MIDI device you intend to use.

Set the response curve, polarity, and range. The response curve may be linear, or exponential, or logarithmic. The polarity is either direct or inverted. To set the Control Range, drag either end of the bar inward. Whenever the bar covers less than the entire controller range, you may also drag it to the right or left, to adjust the range offset.

You may, if you wish, assign the same external control to several TimewARP 2600 sliders or knobs, and set a different curve, polarity, and range for each one. In this way, you can create patches and confi gurations of the TimewARP 2600 that are specifi cally adapted for expressive performance either live or in the studio.

2.2.1.2 Patch-Specifi c Settings

Besides globally assigning a slider, knob or switch to a MIDI controller, you may also assign the slider or knob to be controlled by the velocity and/or aftertouch parameters of incoming keyboard events. (These assignments are not global; they are stored with not global; they are stored with notindividual patches.)

In real time, these controller values are summed with the current global control values to determine the real-time value of the slider. So, by careful tuning beforehand, you can have both keyboard expression and hardware knob/slider control at the same time.

2.2.2 MIDI Patch Selection

In each group of patches, the fi rst 128 are available to standard incoming MIDI patch selection commands.

Patch categories don’t affect this numbering. For example, if the fi rst category in a group has 127 patches, then the second category will have just one patch available for MIDI patch selection.

Groups themselves can be selected with MIDI bank-select commands

This chapter is about the facts – physical, mathematical, and auditory – that make the TimewARP 2600, and the hardware that it emulates, possible. We have to spend a few minutes here distinguishing between physical signals, and the sounds that people hear in the presence of certain kinds of signals.

This is important because synthesizer equipment can only deal with signals – physical commotion of one sort and another. When you are fi ddling with synthesizer equipment, you are generating and modifying signals for the sake of the interesting (we hope) sounds you hear when those signals reach your eardrums.

3.1 Signals and Sounds

A signal is something happening: a waving fl ashlight, ambulance siren, referee fl ag dropping, winking at a friend. Tiny disturbances of the air around us are signals for our ears; we hear them as noise, or singing, or sirens, shrieks, growls, whatever.

The signal is the physical disturbance in the air, the movement of the eardrum. The soundis your perception of the signal: “Hello!”

3.1.1 Analog and Digital Representations of Signals

The signals we are concerned with in sound synthesis are audio signals: more or less regular variations in air pressure, at our eardrums, repeating at rates of between 20 and 20,000 times in one second.

Such signals are straightforward physical processes which can be recorded and reproduced. One way to do that is to look at the pattern of air-pressure variation, and model it in some other medium. During the past century this has been done with grooves in a phonograph record, magnetic fi elds along a length of tape or wire, and other media. The usual scenario is: with one or more microphones, generate an electronic model of the vibrating air, then use the electronic signal to drive a magnetic recording head, or amplifi er, or LP recording lathe. Throughout such processing, the signals we deal with are directly analogous to each other; except for the change in medium from air to voltage to magnetic fi eld strength or stylus position, the signals are identical. Graphed or charted, they even look the same. This is analog recording.analog recording.analog

The Craft of Audio Synthesis 3

12

CHAPTER 3 - The Craft of audio Synthesis CHAPTER 3 -

Another way to record such a signal is, with high-speed digital circuits, to measure the underlying medium many times each second, and store the measured numbers. This is digital recording.

Since the TimewARP 2600 is not constructed of electronic circuits, concepts such as “voltage” don’t apply here. The TimewARP 2600 is software, a complex piece of computer behavior. The signal medium for the original ARP 2600 was electrical pressure, measured in volts. The signal medium for the TimewARP 2600 is simply number sequences. We use those numbers the way the original machine used electrical pressure; where the original Owner’s Manual used the word “voltage” we will just say “signal”

o r “signal level”, and in the module specifi cations we will refer to “virtual Volts” or “vV”.

3.2 Attributes of Signals

The simplest possible signal is a sine wave. It’s like the back-and-forth motion of a point on a circle as the circle rotates. A lot of the mathematics of sine waves is based on that rotating-circle idea; you don’t have to get involved in that unless you’re curious about it, but it’s helpful to train your imagination by picturing the basic sine-wave graph as a slightly stretched coil spring like a “slinky”.

The motion of a pendulum, or of a tuning fork, swinging back and forth as they slowly come to rest, is a decaying sine wave.decaying sine wave.decaying

A sine wave signal has exactly three attributes: its frequency, its amplitude, and its phase. It has no other characteristics at all. (The decaying swing of the pendulum or the tuning fork does have one other attribute: the amount of energy/amplitude that it loses on each swing. It’s not a “pure” sine wave.)

3.2.1 Fundamental Attributes

Picture a point on that rotating circle, leaving a trail behind as it rotates, like an airplane propeller. Picture the trail, stretched out behind like a coil spring, and ask yourself:

Analog signals @ 250 Hz(approximately middle C)

A

B

C

D

E

F

G

H

I

forth

back

forth

back

in

out

higher

lower

higher

lower

up

down

north

south

positive

negative

forth

back

Source

Air PressureWave

MicrophoneDiaphram

MicrophoneOutput Signal

(Voltage)

PhonographLP groove

MagneticField on Tape

Power ampOutput

(Volt * current = Power)

LoudspeakerCone

Your Eardrum

1250 4 msec=

amplitude

etc.

period

frequency = period1( )

phase (important only in comparing two signals)

CHAPTER 3 -

13

CHAPTER 3 - The Craft of audio Synthesis

3.2.1.1 Amplitude: what’s the diameter of this imaginary circle?Amplitude: what’s the diameter of this imaginary circle?Amplitude:

3.2.1.2 Frequency: how fast is it rotating?

3.2.1.3 Phase: when does it start a new cycle?

3.2.2 Complex Attributes

Most of the activity that reaches our ears every day is far more complex than just a sine wave. Banging on a garbage can creates a much more complicated sort of vibration than tapping a tuning fork.

But any kind of motion that isn’t a sine wave can still be isn’t a sine wave can still be isn’tanalyzed as a collection of sine waves. This is called Fourier analysis, after the man who discovered that it was possible, and proved – mathematically – that it always worked.

This goes both ways: any complex signal can also be constructed from a collection of sine waves with the constructed from a collection of sine waves with the constructedappropriate attributes: amplitude, frequency, and phase relationships.

In fact, any kind of motion – any repeating pattern of the sort that we are interested in here as “sound waves” - can be examined and analyzed under two different headings:

3.2.2.1 Waveshape, and Time-Domain Attributes

Any repeating motion can be diagrammed in a graph where the horizontal axis is a period of time and the vertical one is the motion itself, back and forth. This is the time-domainview of a signal:

3.2.2.2 Spectrum, and Frequency-Domain Attributes

Instead of looking at a signal as something in motion through a period of time, we can look at the collection of sine-wave components of the signal. In such graphs, the horizontal axis is a frequency range (instead of a time period), and we indicate each spectrum component with a single vertical line in the graph. The height of the line indicates the strength of the component at that frequency. This is the frequency-domain , or spectral-domain, view of a signal.

It is the distribution and relative strength of spectral components that we experience as the tone-color of a sound or sounds. “Bright”, “dull”, “sharp”, “tinny”, “heavy”, and so on – these are all descriptive words for the spectral attributes of sounds.

A complex signal...

can be made of simple ones like this...

...and this...

...and a pinch of this

...and a dash of that.

14


3.2.2.2.1 Harmonic Series

The spectral view of any periodic signal has components at simple multiples of the signal frequency. For example, suppose we examine a sawtooth wave at a frequency of 110Hz. It will have components at 110, 220, 330, 440, and so on.

A simple number sequence like this is called a harmonic series. It is interesting to think about musically; the smaller numbers in the series all form simple musical intervals: octaves, fi fths, fourths, and thirds.

Most musical instruments generate harmonic spectra. Some have more, some have fewer harmonics, and there are wide spectral variations even within a single instrument depending on how it is played. In general, whatever the actual spectral components are, they will always form a harmonic series.

In audio synthesis, you will use oscillators to generate pitched tones that have a harmonic spectrum.

Spectrum = the sine-wave component of a signal

Incr

easi

ng

Po

wer

Increasing Pitch

1 2 3 4 5 6 7 8 9 10

@250Hz

sinewave

sawtooth

guitarstring immediately

after pitch

just beforedying out

1/8pulsewave

(octaves)

middle C 250Hz

A harmonic series is composed of numerically equal frequency intervals, and that means decreasing pitch intervals

Incr

easi

ng

Po

wer

Increasing FrequencyN 2N 4N 8N 16N 32N 64 128 256 512 1024

1 2 3 4 5 6 7 8 9101112

The first couple of octaves’ worth of a harmonic series make useful musical intervals: octaves, 5ths, 4ths, 3rds and so on. Beyond the 16th, they get too close together to sound good.

Incr

easi

ng

Po

wer

Increasing PitchANY ANY*2 *3 *4 *8 *16

octave

8

5th

maj3rd

min3rd

Pick any frequency; the multiples of that are a harmonicseries. A, B, and C are all harmonic series.

Incr

easi

ng

Po

wer

Increasing Pitch

C

B

A

etc.

2Khz 4K 6K 8K

etc.

100Khz 200

250Hz 500

12501500

1750

750 1K 2K

300 400

CHAPTER 3 -

15


3.2.2.2.2 Enharmonic Series

Enharmonic, in this context, means “not forming a harmonic series.” Some waveforms do not repeat themselves at a regular interval; the spectra of such waves will have components at strange non-integral frequency ratios, or they may have shifting spectral components that die out and reappear. Some percussion instruments, such as kettledrums, or bells, have enharmonic spectral components.

In audio synthesis, you can use various modulation techniques – such as AM and FM - to generate enharmonic spectra.

3.2.3 Attributes of Random Signals

A completely random signal – noise – actually has a continuous spectrum: it does not consist of isolated sines in a harmonic series, nor even in an enharmonic series. A noise spectrum is a continuum of frequency components; in order to describe it, we have to talk about how much energy is in each band in this continuum. One kind of noise may have high-frequency energy, and another has low-frequency energy.

In audio synthesis, noise is an extraordinarily useful signal. Filters can be used to shape a noise spectrum into almost anything – even pitched sounds.

3.2.3.1 Spectral Balance (Color)

The physical processes (such as molecular motion or random popping of electrons from one atom to another) that we typically depend on for random electronic noise have a frequency distribution of equal probability at any frequency, or through any frequency interval. Curiously enough, this produces noise signals that, for audio purposes, don’t sound properly balanced across our range of hearing. They sound too bright.

Not all spectra are harmonic. An Arbitrary collection ofsines at unrelated frequencies is "unharmonic".

Incr

easi

ng

Po

wer

Increasing PitchANY

etc.

+1.7

1.47

2.43

2.725

3.157

16


3.3 Attributes of Auditory Events

Of all the thousands of sounds you hear in a single day, only some have a defi nite beginning or ending point. Many just come and go without a clear start/stop. Those that do have a reasonably clear start/stop we will call auditory events. The other sounds, the less clearly defi ned ones, we can call auditory textures or backgrounds or even structures.

Think how some sounds are more complicated than others. The hum of a refrigerator is sort of simple, the hum of a tuning fork is (as we heard above) really simple. A single bass note from a piano is quite complicated, the way it keeps changing and evolving as it dies away. Human speech is a very complicated kind of sound, even when it’s your native tongue (when it’s a language you don’t understand, it sounds even more complicated).

There are many, many more kinds of sounds to hear than you or I have words to describe them with. Here are just a couple of the most general distinctions people make about different kinds of sounds:

3.3.1 Steady-State Attributes: some sounds, once they get started, remain pretty constant. They don’t change much while they continue, and when they stop, they just stop. We say such sounds reach a steady state, in which we can pick out a defi nite steady state, in which we can pick out a defi nite steady Pitch, Loudness, and Tone Color. Notes from organ pipes – or from electronic drawbar organs – are steady-state sounds.

3.3.2 Time-varying Attributes (i.e. “Envelopes”): some sounds have a pretty clear start and end, but while they’re happening they change. Think of a bird song, or ordinary human speech.

Such sounds can sometimes be analyzed as more or less rapidly changing in one or more of those fundamental auditory attributes: pitch, loudness, or tone-color. Think, for example, of how the sound of a plucked guitar string evolves from the moment you pick it to the moment you can’t hear it anymore. It starts out loud and fades away; and it starts out “bright” - with lots of harmonics – and is slowly “muffl ed” as it fades out.

One way to conveniently diagram what happens in such events is to chart each changing attribute separately, in its own time-domain graph.

Such a graph is often referred to as an envelope. Over the years, some standard vocabulary has developed for talking about envelopes: Attack (for the fi rst part of an event), Decay (for Decay (for Decaysome later parts), and Release (for the last part, when the event is coming to an end).

“white” noise has equal energy in any two equal frequency bands…

0dB

-30dB

so it’s NOT equal in equal pitch intervals…

…even when you use dB for the vertical axis…

THIS is why white noise sounds so shrill

octave intervals

20 4

-12-9dB -6 -3

6 8 10 12 14 16 18 20K

20 40 80 160 320 640 1.25 2.5 5 10K 20K

CHAPTER 3 -

17


3.4 How Signals and Sounds Go Together... Sort Of

Signal activities, being entirely physical kinds of things, are easily nameable, measurable, catalogable, countable. Sounds, as we pointed out above, are not quite so easily domesticated. The music and other sounds that we listen to correlate in several well-established ways with the signals around us; but the correspondence is not simple. There are always surprises.

3.4.1 Signal Frequency and Audible Pitch

This is probably the best-established and most reliable correspondence. It dates all the way back to Pythagoras, the Greek philosopher who 2500 years ago worked out that halving the length of a vibrating string made the pitch rise by one octave.

3.4.1.1 How Adding Pitches Means Multiplying Frequencies

The range of audio frequencies – of human hearing, in other words – is conventionally stated to be 20Hz to 20KHz. And when you think linearly, it sounds tragic to learn of someone, say, who can only hear up to 10KHz. But in the realm of human pitch perception, such a person has kept 90% of his hearing range: from 10KHz to 20KHz is just one octave out of the 10 we hear.

This is a really important fact in really important fact in really important factaudio synthesis; you will encounter it over and over again, in every patch you create. It governs much of the arithmetic of generating and controlling spectral distributions and patches:

4 Equal pitch intervals require exponentially increasing frequency increments;

4 Equal frequency increments make a harmonic series. In a harmonic series, as we go frequency increments make a harmonic series. In a harmonic series, as we go frequencyup the series we get smaller and smaller pitch intervals.

A harmonic series is composed of numerically equal frequency intervals, and that means decreasing pitch intervals

Pitch by octaves

Freq

uen

cy in

cyc

les

/ se

con

ds

0 1 2 3 6 7 8 9 10middle C

Each divisionis one octave.To add an octave, you haveto multiply F x 2.

20Hz 40Hz 78Hz 156Hz312Hz 625Hz

1250Hz2500Hz

5KHz

10KHz

18


3.4.2 Signal Amplitude and Audible Volume

This is a very dicey relationship. It is true that, for any given signal, increasing its amplitude will increase the volume of the associated sound. But a lot of other signal characteristics have a greater impact on our perception of volume than amplitude does.

For example, the difference between talking and shouting at someone is far more a matter of pitch and spectrum – tone-color – than of mere amplitude. When I yell, I “raise my voice”; that is, I raise the pitch of my voice, and I put more energy into it, which generates more harmonics, which is a matter of spectral content, not amplitude. Stage actors have to learn to overcome this tendency; in order to be heard onstage without shouting, they have to learn to increase their speaking amplitude without yelling.

If you examine – visually - the recorded signal from a pipe organ, or even an orchestra, you may be surprised to fi nd the apparent amplitude of the softer signals almost equal to that of the loudest. This has to do with the spectral distribution of the sound. Of two different signals of approximately equal amplitude, the one with the broadest spectral distribution – the most harmonic content - will sound louder.

3.4.2.1 How Adding Volumes Means Multiplying Amplitudes

Nonetheless, for a given signal and its spectrum, it is always true that a bare increase in amplitude will be heard as an increase in loudness. But how much of an increase?

It turns out that, just as with frequency and pitch, the relationship here is exponential. A century of research has established that equal steps in loudness are represented by multiplicative ratios in signal amplitude.

In other words, if you double the amplitude of a signal, you will hear an increase in loudness. But then to get another increase equal to the fi rst one, you have to double the amplitude again.

3.4.3 Signal Spectrum and Audible Tone-Color

This is a quite solid relationship. In fact, the word “spectrum” has achieved a meaning in both worlds: depending on the context, it can refer to a measurable attribute of physical signals, or a character of perceived sounds.

Any change you hear in the character of a sound – in its tone-color or “spectrum” - must have an associated variation in the character of the physical signal that is arriving at your eardrums. Tone change = waveform change.

The reverse is not quite so certain. It’s fairly easy to fi nd waveform changes that listeners can’t hear. For example, we humans simply aren’t sensitive to phase relationships within a complex spectrum. But, in the time domain, two identical spectra with shifted phase relationships among their components can look unrecognizably different.

A 3dB rise or fall in sound level is noticeable, but it's not much; it represents a factor of 2 change in signal power.

-33 -30 -27 -24 -15dB-21 -18 -12dB -6dB Ref-3dB ReferenceSound level

ReferencePower level

decreasing sound level

(any)

decreasing Power level

etc.

Power8

Power4

Power2

CHAPTER 3 -

19


3.4.4 Signal Envelopes and Audible Event-Contours

One of the most fascinating areas of audio synthesis is listening for the envelopes of time-varying events. Here there are all sorts of mysteries, in which signal attributes get regularly “misperceived”: frequency variations get heard as volume, spectral evolutions are heard as pitch, and amplitude envelopes generate an elusive spectral twitter.

3.5 Modules and Methods for Generating Signals

How can you get access to the signals and processes we’ve just described, so that you can play with them on your own?

Throughout the past century people have been noodling around with electronic ways of generating audio signals. In particular, beginning in the 1960’s, people such as Bob Moog, Don Buchla, and Alan R. Pearlman began to settle on some ideas that have become almost standard for audio synthesis: independent, modular functions for signal generation independent, modular functions for signal generation independent, modularand signal processing, capable of being controlled not only by hand but also by signals of the same kind as they generate.

This was the idea of voltage-controlled operation, and it was completely revolutionary. voltage-controlled operation, and it was completely revolutionary. voltage-controlledUsing independent, modular functions made it possible to change one attribute of a signal without necessarily affecting any other attribute; so the craft of synthesis became the craft of constructing, and tuning, integrated confi gurations of modules.

The cables that connected modules were called patchcords, and so connecting modules together came to referred to as patching them, and so, fi nally, any working confi guration patching them, and so, fi nally, any working confi guration patchingcame to be called a patch.

Let’s take a look at some modules that generate signals.

3.5.1 Oscillators

A device that repeats the same motion over and over is an oscillator. In audio synthesis, oscillators typically produce very simple geometrical-pattern signals such as sine, triangle, pulse, and sawtooth waves, named simply for what their time-domain graphs look like.

In an analog synthesizer, the analog synthesizer, the analogunderlying medium in motion is usually electrical pressure, or voltage. In a digital synthesizer, digital synthesizer, digitalthe signal is actually generated as a sequence of calculated numbers. (Conventionally, these are generated at the standard sampling rate for music CDs, 44.1KHz.) This does not become motion until, at the output of the synthesizer, the number stream is converted to variations in voltage, and then amplifi ed, and then used to drive a loudspeaker. (A loudspeaker is a motor that moves back and forth instead of around and around.)

Because oscillator-generated signals are periodic, their spectral components always form a harmonic series.

9

Sawtooth-3dB / Octave

Pulse

Triangle

Sine

Oscillator Output

TIME - domain

SPECTRUM

FREQUENCY - domain

t

f 2f 3f 4f 5

f 3f 5f

10 136 7 8f 16f

f 2f 4f 8f 12f 16f

1/t

OdB

-10

-20

OdB

OdB

-10

-20

-10

-20

OdB

20


3.5.2 Noise Generators

A device that jiggles at random without ever repeating itself is a noise generator. Waterfalls, steam, wind, fans, and such things are all noise generators.

The spectrum of a noise signal is a statistical distribution of frequency components. (This is the opposite of a sine wave, which is exactly one frequency.) A noise spectrum that is perfectly balanced throughout the musical range is called pink noise. Pink noise is very useful in listening tests of loudspeakers, because a trained human listener can hear even tiny differences between two different noise spectra.

Filtering and equalization can shape a noise spectrum into almost any sound.

3.5.3 Envelope Generators

A device whose output is intended to control some time-varying attribute of an event is called an envelope generator. These are sometimes referred to as transient generators, to call attention to the fact that their output is not constant but transient.

An envelope generator produces an output signal only “on demand”. The demand is made by means of timing signals called gates, and triggers.

3.5.4 Sample & Hold Processors

The idea of “sampling” a signal does not directly relate to any particular characteristic of audio events; instead, it is an idea from electronics that has turned out to be useful for creating patterned control signals.

3.6 Modules and Methods for Processing/Modifying Signals

3.6.1 Inverters

A signal inverter works exactly like a seesaw. When the input goes high, the output goes low, and vice versa. An analog inverter would output negative voltages on positive input; a digital inverter simply multiplies its input number-stream by (-1).

Whitenoise

spectrum

Pinknoise

spectrum

Red (LF)noise

spectrum

0dB

0dB

0dB

+15

-15

-15dB

-30dB1 2 3 4 5 6 7 8 9 10

gate

trigger

ADSRenvelope

ARenvelope

envelope generators are controlled by gateand trigger signals...

CHAPTER 3 -

21


3.6.2 Signal Mixing

A signal mixer adds two or more signals together and outputs the result of the addition. This is a more complex signal, usually, than any of the inputs. But not necessarily; if signal B, for example, is the exact inversion of signal A, then mixing the two will produce a signal of exactly zero.

3.6.3 Attenuators/Amplifi ers

An attenuator cuts the strength of a signal passing through. Digitally, this is accomplished by multiplying the input by some value ranging from 0.0 (which passes no signal) to 1.0 (which passes the signal at its full input amplitude).

An amplifi er may increase the amplitude of a signal. But not necessarily; it is usual for a voltage-controlled amplifi er to have a maximum gain factor of 1.0. In fact, the purpose of VCA’s is actually to “chop” the signals passing through them. It would make more sense to think of them as “voltage controlled attenuators”.

3.6.3.1 Gain Factor

To describe the behavior of an amplifi er or attenuator, we may use the expression “gain factor” to mean the ratio of output signal amplitude to input amplitude.

3.6.4 Filters

A fi lter is a device that works better at some frequencies than at others. (The inverters, mixers, and attenuators we have been describing work the same at all frequencies, so they are not fi lters.)

Because of this frequency-dependent characteristic of fi lters, they change the shape of any complex waveform passing through. And so it will be important for you to get to know what fi lters do to signals in both the time domain and in the frequency domain.

Inverters

input

input

output

output

these foursignals added(mixed)together make....

two sines added together

22


3.6.4.1 Low-Pass Filters

Any device or mechanism that passes along slower motions better than faster ones can act as a lowpass fi lter.

Picture yourself stirring a cup of tea with one of those little wooden paddles they hand out in the coffee shops. Stir it back and forth, fast. Now slow down. Now imagine the tea has turned to syrup. You can still stir it slowly, but if you try to go fast the stick will simply not move.

That’s a lowpass fi lter. You can see the effect of this on a signal quite easily.

For audio signals, you will usually be more interested in the frequency-domain effects of fi ltering. For subaudio signals, it is usually the time-domain effects – changes in waveshape – that we care about. In the time domain, a low-pass fi lter rounds off

any sharp transitions in the signal. A good example of this is the lag processor described lag processor described lag processorin section 3.6.4.5 below.

In the frequency domain, it weakens spectral components that are higher than the fi lter cutoff frequency – the frequency at which the fi lter begins to have an effect on the signal.cutoff frequency – the frequency at which the fi lter begins to have an effect on the signal.cutoff frequency

3.6.4.2 High-Pass Filters

Any device or mechanism that passes along faster motions better than slower ones is a highpass fi lter.

Take the drinking straw from your water glass. Seal the end of it with your thumb and dip it back into the water. Notice that you can pump it up and down in the glass, fast or slow, but the water never leaks into the straw as long as you hold your thumb over the end. Think of the up and down motion of the water at the bottom end of the straw as “the signal”.

Now start letting a little air leak into the straw as you move it up and down. The water level at the bottom end of the straw no longer stays down when the straw “signal” goes down – it starts to come up again. And then when you draw the straw back up, the water leaks back down more or less rapidly depending on the position of your thumb at the top of the straw.

This is a high pass fi lter. In the time domain, it constantly “leaks” its output signal level back to zero, at a rate related to the cutoff frequency and slope. In the frequency domain, it passes all spectral components higher than the cutoff frequency, and attenuates those below the cutoff frequency by an amount proportional to the cutoff slope.

A lowpass filter inthe time domain . . .

f 4f 8f 16f 32f| | | |

. . . makes it difficultfor the medium tochange value.

. . . and in the frequencydomain, this cuts offhigher-frequencycomponents.

T=1/f f 2 4 8 16 32

A highpass filter makesit difficult for the mediumto keep a new level - it"leaks" away . . .

. . . and so lower-frequency componentsare weakened.

CHAPTER 3 -

23


3.6.4.3 Cutoff Slope

This is the rate at which a fi lter attenuates spectral components, as a function of their frequency. It is usually a multiple of 6dB/octave.

3.6.4.4 Feedback and Resonance

It is usually possible, with any signal-processing device or system of devices, to mix some of its output signal back into the input signal. This may be intentional, or it can happen by accident; everybody who has ever worked with a PA system has experienced the terrible screech of a system “in feedback”.

In fi lter modules for audio synthesis, it is common to provide controllable feedback. With just a little, the fi lter response begins to peak around its cutoff frequency; as the feedback level increases, the peak gets stronger. Eventually – just as happens with a PA system – the fi lter falls into oscillation. Regardless of the input signal, it “screams” a sine-wave at its cutoff frequency. In this state it is no longer behaving as a fi lter at all; it has become an oscillator.

Systems in feedback have been very well studied in physics. Their behavior can be described mathematically. Whereas feedback in a PA system can be unpredictable and uncontrollable, feedback in a fi lter module for audio synthesis can be – and is – a controllable and useful feature.

3.6.4.5 Lag Processors

A lag processor is a low-pass fi lter intended specifi cally for processing subaudio signals. It lag processor is a low-pass fi lter intended specifi cally for processing subaudio signals. It lag processorintroduces a “lag” in the output signal wherever the input shows a sharp change in value. How much of a lag depends on the “cutoff frequency” of the processor.

3.6.5 Modulation Methods

The simplest possible signal, we said above, has exactly three attributes and no more: frequency, amplitude, and phase. If, starting from a steady-state signal, we systematically modify any of these characteristics, we are said to be modulating the signal. And so, based on these three signal attributes, there are three possible forms of modulation: Amplitude Modulation (AM), Frequency Modulation (FM), and Phase Modulation (PM). The fi rst two are more commonly used in audio synthesis than the third; we won’t say anything here about phase modulation.

Using AM and FM methods, it is possible to generate waveforms and spectra that are far more complicated – and interesting to the ear – than anything that can be produced by merely mixing and fi ltering signals. Mainly that is because of sidebands.

24


3.6.5.1 Sidebands and Sideband Spectra

What happens to the spectrum of a sine wave when we modulate its amplitude? What happens when we modulate its frequency?

Clearly, since the signal that results from AM or FM methods is no longer a plain vanilla sine wave, then, in the frequency domain, it must have some additional components. These additional components are called sidebands. They have been studied for at least a century, and are pretty well understood physically and mathematically.

Audio synthesis is probably the only application of AM or FM modulation where we are interested in sidebands for their own sake, as something to listen to directly; in the past, this stuff was only interesting to radio engineers, radar, sonar, television broadcast engineering. In those disciplines, modulation sidebands are products of broadcasting methods, in the electromagnetic spectrum. Because of that historical background, some of the conventional language for talking about modulation processes is a little weird: the signal being modulated is often referred to as the carrier signal, and the signal that carrier signal, and the signal that carrier signalprovides the modulation pattern is called the program. (Guess why.)

Any form of modulation generates, for each component of the original signal, at least one lower sideband – at a frequency equal lower sideband – at a frequency equal lower sidebandto the component minus the modulating frequency – and at least one upper sideband, at a frequency equal to the component plus the modulating frequency. If the carrier – the original signal - is itself complex, with multiple spectral components, then each of its components will produce its own sidebands independently of all the others. Likewise, if the modulating signal – the program - is complex, the arithmetic applies separately to each of its components.

So, just for example, if you modulate a 10-component carrier signal with a 10-component program signal, the signal resulting from the modulation will have not less than 100 spectral components. This can get very messy; the most useful thing you can do with such a signal, before you do anything else, is fi lter it to get rid of some of the fuzz.

3.6.5.2 Amplitude Modulation

Suppose we modulate the amplitude of a 1000Hz sine wave with a 5Hz sine wave. The result is indistinguishable from what we would get if we mixed three sine waves, at 995Hz, 1KHz, and 1005Hz. They are the same signal.

The 995Hz component of the output is the lower sideband resulting from the modulation, and the 1005Hz component is the upper sideband.

Given this for a carrier:

. . . and this for a program

here's the AM result:

carrierspectrum

programspectrum

uppersideband

lowersideband

carrierweakend

the carrieris gone

and varyingamplitudedependingon the modulationdepth

?

variablenumber ofsidebands

here's the ring modulation:

and here's the FM result:

CHAPTER 3 -

25


3.6.5.2.1 Ring Modulation

Whereas a VCA responds only to a positive-going signal at its amplitude-control input, a ring modulator responds to both positive and negative levels at both of its inputs. Its output is simply the product, arithmetically, of the two inputs. If you are new to audio synthesis, draw a couple of signal graphs – it doesn’t matter what they are – on the same timebase and vertical scale, and use a pocket calculator to work out the result of multiplying the two signals together. That’s what a ring modulator does. (The expression “ring modulator” describes the appearance of the analogue circuit design that’s required for the multiplication.)

In the frequency domain, the difference between this and ordinary AM is only that the carrier signal components are suppressed. Once again, if you work out the arithmetic, a single-frequency carrier, modulated by a single-component program, generates a three-component AM spectrum but only a two-component ring-modulation spectrum.

What’s useful about this? Well, since the carrier is almost always periodic (it comes from an oscillator, right?), it has a harmonic spectrum. Suppressing this spectrum lets you hear Suppressing this spectrum lets you hear Suppressingjust the sidebands, which can be completely enharmonic if you’re careful about the ratio of the two input signal frequencies.

3.6.5.2.2 Frequency Shifting

It’s theoretically possible not only to suppress the original carrier, as in ring modulation, but to isolate the lower and upper sidebands and make them available separately. The arithmetic here is fascinating, because the end result (for once) is in one-to-one correspondence with the input: for each component of the program signal, there is a component in the output at C-p (or at C+p). In other words, the fi nal spectrum has only as many components as the original program did. This is called frequency shifting. Picture the entire program signal spectrum shifted up or down by some fi xed frequency.

The important thing to remember about this is that it’s not pitch shifting – which would have to be accomplished by frequency multiplication – but frequency shifting. It’s an frequency shifting. It’s an frequencyaddition or subtraction process, and it really messes up any harmonic relationships that really messes up any harmonic relationships that reallymight have existed in the original spectrum.

3.6.5.3 Frequency Modulation

The spectrum resulting from amplitude modulation always has three components for every one component of the program signal: the carrier itself, and two sidebands. In Frequency Modulation, however, the number of sidebands depends on the modulation depth. It is possible from only two sine waves to generate a spectrum with dozens or even hundreds of components. Modulating one sawtooth with another can produce a spectrum so complex that it sounds almost like a noise generator. In such a patch, you will usually reach for a fi lter to take the edge off the resulting spectrum.

What happens is this: as the depth of modulation increases, the number of sidebands does too, without limit. The additional sidebands come in at – guess what – integral multiples of the program frequency.

For this reason, the most useful FM techniques involve only sine-wave carrier and program signals.

26


3.7 Controlling One Module by Means of Another

The modulation methods we’ve just described can be accomplished with voltage-controlled or digitally-controlled equipment. For example, to set up an AM effect, feed the carrier signal into a VCA audio input, and the program signal into one of the amplitude-control inputs.

Likewise, to set up an FM method, route the program signal into one of the frequency-control inputs of a VCO. (See section 4.3 for news of some TimewARP 2600 extensions relating to this.)

3.7.1 Linear and Exponential Sensitivity To Control Signals

In designing and constructing a voltage-controlled device, or a digitally-controlled algorithm, we have to consider how we intend the controlling parameter to relate to the controlled parameter.

Suppose, for example, we build an oscillator that changes its output frequency by exactly 1000Hz for each rise of 1 volt in electrical pressure at a designated point (a “control input”) in the oscillator circuitry. This is a linear relationship between the control value and the frequency. In order to march such an oscillator through a musical scale, we would have to provide exponentially increasing voltage steps for each rising musical interval. That’s pretty awkward, and it gets worse very rapidly. The pitch interval produced by, say, a 0.5-volt control step will depend – completely – on the initial frequency of the oscillator.

Suppose we tune this imaginary linear-responding VCO to 500Hz, and apply a sequence of 0.5V control steps. What frequency do we get at each step? What musical interval?

And so, instead of linear sensitivity, VCO’s for audio synthesis are designed, almost universally, with exponential sensitivity to control signals. Such a design sets up an exponential sensitivity to control signals. Such a design sets up an exponentialexponential relationship between control signal and frequency in order to maintain a simple linear relation between control signal values and audible pitch changes.

Similar arguments apply in the design of voltage-controlled fi lters: it’s more practical to work with a linear relationship between control-signal and “cutoff timbre”, and therefore the relationship between control-signal and cutoff frequency really has to be exponential.

0

+0.5V

+1.0V

+1.5V

+2.0V

+2.5V

500Hz

1KHz

1.5KHz

2.0KHz

2.5KHz

3.0KHz

unison

octave

Fifth

fourth

Major third

Minor third

INITIAL CV FREQUENCY INTERVAL

CHAPTER 3 -

4.1 Top Row Control Panel Buttons and Indicators

Just above the panel graphics, outside the “case” of the TimewARP 2600, is a horizontal row of buttons and indicators for patch storage, import/export, voice-cloning, and other operations.

These powerful features of the TimewARP 2600 have no equivalent in the world of analog synthesis; they are unique digital extensions of the original ARP 2600 synthesizer.

4.1.1 Patch Lock Button (Padlock Icon)

This padlock button helps you to avoid accidentally overwriting your favorite patch. Basically it disables the Save button, while leaving the Save As button enabled. Under these conditions, you can save your current patch only by assigning it a new name.

4.1.2 Group, Category, and Patch Drop-Down Lists

The TimewARP 2600 gives you a three-level hierarchy for storing and organizing your patches. All Patches are sorted into various Categories, which are in turn sorted into major Groups. Each of the three patch selection buttons generates a drop-down list associated with one layer in this hierarchy.

Groups, Categories, and Patches can also be selected by keyboard shortcuts. The up/down arrow keys on the computer keyboard select Patches, the left/right arrow keys move between Categories, and using the control key with the left/right arrow keys moves between Groups.

4.1.3 Save Button.

The Save button saves the current patch confi guration and settings under the name of the most recently loaded patch. This button is disabled if the patch is locked (see 4.1.1 above).

4Modular Components of the TimewARP 2600

28

CHAPTER 4 - Modular Components of the TimewARP 2600 CHAPTER 4 -

4.1.4 Save As Button.

The Save As button saves the current patch confi guration and settings under a group, category, and patch name of your choice.

Within the Save As dialog, you may create new groups and categories at will.

There is no limit to the number of groups and categories you may create.

4.1.5 Patch Manager Button

Use the Patch Manager to organize Patch Manager to organize Patch Managerand reorganize your patches; to move patches from one category to another, and to move whole categories from one group to another. Use it to export and import patch collections – dozens or hundreds of patches at a time.

The Patch Manager window displays Patch Manager window displays Patch Managerall three levels of the patch hierarchy, and supplies a number of tools for managing the entire collection.

These tools are listed in a column on the left of the window.

To use them, select one or more items from the hierarchy, and then click on the operation you want to perform. Any operation that cannot be applied to the current selection of items will be disabled in the list.

CHAPTER 4 -

29

CHAPTER 4 - Modular Components of the TimewARP 2600

4.1.5.1 Import / Export

Use the Import / Export commands to write or read entire Import / Export commands to write or read entire Import / Export Groups, Categories, and Patchesto/from external fi les.

Export works on whatever items are currently selected; by selecting all of the current Export works on whatever items are currently selected; by selecting all of the current Exportgroups, you can use Export to make a backup of all of your patches. Export to make a backup of all of your patches. Export

When you export a single Patch, the names of the Group and Category for the patch are Category for the patch are Categoryexported with it.

Import asks you to select the fi le you want to read in. Import asks you to select the fi le you want to read in. Import Import will never overwrite any Import will never overwrite any Importexisting Group, Category, or Patch; if any Group or Category or Category or Category Patch in the fi le to be imported has the same name as a Group or Category or Category or Category Patch that already exists, Importwill append a number to the loaded name. So when a friend sends you his collection of a thousand patches, you can import them without worrying about possible name overlaps.

4.1.5.2 Cut / Copy / Paste

The Cut / Copy / Paste buttons appear within the Patch Manager dialog to move items from Patch Manager dialog to move items from Patch Managerone place in the hierarchy to another. You can, for example, Cut a patch, or a group of Cut a patch, or a group of Cutpatches, from one category and Paste it into another.

4.1.5.3 Up / Down

The Up / Down buttons appear within the Patch Manager dialog to move selected items Patch Manager dialog to move selected items Patch Manager Upor Down in their list. (This is useful for arranging your MIDI patch lists.)

4.1.5.4 Rename / Delete

The Rename / Delete buttons appear within the Patch Manager dialog to Patch Manager dialog to Patch Manager Rename or Delete selected items.

4.1.6 Voice Button

Clicking on the Voice button activates a drop-down list from which you may select the number of simultaneously sounding voices you want to use.

Because the TimewARP 2600 is a true analog synthesizer emulator, its modules are running even when no audio signals appear at the synthesizer’s output. Each voice added to the multi-voice capability of the TimewARP 2600 is a clone of the entire patch and module set. This will have an immediate and obvious effect on the CPU load meter.

So: how many voices you can generate without overtaxing your CPU will depend on your machine’s clock speed.

30


4.1.7 Reset Button

The Reset button removes all patch cords and returns all sliders to a standard position.Reset button removes all patch cords and returns all sliders to a standard position.Reset

4.1.8 MIDI Indicator

This virtual LED glows when there is any MIDI input to the TimewARP 2600 – not just keystrokes, but also controller input and sysex dumps.

4.1.9 Output-Level Meter

This shows the output signal level. If it reaches into the red segment, your signal will distort.

4.1.10 CPU Load Meter

This meter shows, roughly, how much of the time between samples (the sample period) is being devoted to the TimewARP 2600 emulation process. In a complex patch, or a many-voice polyphonic performance, the meter may indicate overload; when this happens, it is likely that the TimewARP 2600 output signal will be interrupted, so your audio feed will develop a glitch. To avoid this, you will have to simplify your patch, or decrease the number of voices, or acquire a faster, more capable computer.

4.1.11 The Magic Logo

At the lower right of the main panel is the TimewARP 2600 Logo. Clicking on this brings up a menu:

4.1.11.1 About TimewARP 2600 identifi es the team; the people who worked together to bring you this software.

4.1.11.2 Load/Save MIDI Maps

Use the Load/Save MIDI Maps commands to save – and reload – the MIDI-controller to slider assignments that you set up. In the TimewARP 2600, these are global assignments, independent of any particular patch settings; saving a patch does not save these assignments, and loading a patch does not change the current assignments. You can, if you want, set your mappings once, and they will be there throughout all of your personal patch changes.

CHAPTER 4 -

31


4.1.11.3 Load Microtuning

You may also load alternate tunings for the keyboard. These are described in Appendix 6.1. The TimewARP 2600 does not allow you to modify these tunings or to save or create new ones.

Microtunings are a global attribute of the keyboard; once loaded, the tuning will govern anything you play until you load a different one, regardless of your patch changes

4.1.11.4

MIDI Beat Synchronization

You may synchronize the Internal Clock (IC) (see section 4.13), to the MIDI Beat Clock (MBC) by specifying the number of MBCpulses per IC transition. As a reference, there are 24 MBC pulses per quarter note.

The keyboard LFO (see section 4.14.1) may also be synchronized to incoming the MBC, independently of the Internal Clock.

Setting different sync counts for these is a fun way to program complex rhythms that are locked to the tempo of your MIDI tracks.

In order for the MBC messages to be sent to the TimewARP 2600, you must enable MIDI Beat Clock in the Pro Tools MIDI Menu, and select the TimewARP 2600 as a recipient of these messages. Also, MBCmessages are only sent when the Pro Tools transport control is running.

MBC synchronization is a MBC synchronization is a MBC patch attribute, not a not a not global one; the sync counts you set here will be stored global one; the sync counts you set here will be stored globalwith the current patch when you save it.

32


4.1.12 Jacks, Patchcords, and Default Connections

The panel has eighty-one mini-jacks. Forty-fi ve are inputs, twenty-nine are outputs, and 7 operate as both input and output.

Of the 45 inputs, 32 are in a row running across the center of the panel. (There are actually 34 jacks in the row, but the two labeled “gate” and “trig” are outputs.) This row of input jacks divides the control surface almost evenly in half.

Above this row, in the upper half of the control surface, there are only three input jacks. input jacks. inputThey are at the upper right corner, labeled Left Input, Pan, and Right Input respectively. All Right Input respectively. All Right Inputof the other jacks in the upper half of the control surface are outputs.

In the lower half of the control surface are inputs to the voltage processors, and of the column of four jacks in the section labeled Sample & Hold, the upper and lower jacks are inputs.

The seven jacks that are both input and output, belong to the Electronic Switch and the Multiple outlet. Because the switch works in either direction, it has either two inputs and one output or one input and two outputs. The Multiple output distributes at least one input to 1, 2, or 3 outputs.

All the remaining jacks are outputs. Most of them are labeled as such; a few are not, but have arrows pointing to them. For example, in the Voltage Processor (VP), the three jacks furthest to the right are outputs; and in the Envelope Generator section of the upper half of the panel, the two jacks labeled Gate and Trigger are outputs.

4.1.13 Sliders and Slider Operations

There are fi fty-eight sliders. Thirty-six of these are plain old signal attenuators. And of those 36, 29 are all in a row across the middle of the panel. (There are actually 31 sliders in the row, but the two on either side of the box labeled Attack Release are not plain old signal not plain old signal notattenuators.)

Most of the attenuators are directly associated with either an input to something or an output from something. Each vertical attenuator across the middle of the panel, for example, adjusts the strength of the signal coming in from the input directly below it.

4.1.14 Normalled Jacks

The most commonly used signal connections are “normalled” (i.e. defaulted). A default signal is identifi ed by a small icon at an input jack.

To see a simple example of this, note that the fi rst input jack on the left, in the row of jacks running across the middle of the TimewARP 2600, is an input to the Envelope Follower. The symbol underneath this jack indicates that the default signal to this input comes from the Preamp. That means that the Preamp output is prewired to the Envelope Follower input, except when a plug is inserted into the jack.Follower input, except when a plug is inserted into the jack.Follower

For another simple example, note that in the same row of jacks, the third one from the right is a mixer input, and that the symbol just beneath it indicates that the default signal to this input comes from the Voltage-Controlled Filter (VCF):

Note again that the fi fth of the fi ve audio inputs to the VCF is similarly defaulted from the VCF is similarly defaulted from the VCFNoise Generator (counting across from left to right this is the 21st jack):Noise Generator (counting across from left to right this is the 21st jack):Noise Generator

CHAPTER 4 -

33


So you can listen to this input by opening the VCF input to the mixer, and the VCF input to the mixer, and the VCF Noise Generator input to the VCF. Now experimenting with the two horizontal control sliders at the top of the VCF panel will give you a wide range of fi ltered sounds.

It will be worth your while to experiment thoroughly and systematically with the default signal connections at this point, particularly if you are planning to use the TimewARP 2600 in live performance. In section 4 we will document the behavior of each separate module, and in section 5 we give sample patches for further experimentation; here we will only mention a few general principles to keep you from going out of your skull with complications:

Experiment with one signal at a time. With the VCF, for example, when you have listened to everything the fi lter can do with a noise input, close that input and open the default VCO-3 sawtooth immediately to its left. Now you can experiment not only with the VCF controls, but also with the manual frequency controls of VCO-3; and when you have done that, experiment one by one with the control input signals to VCO-3.

4.2 Preamp/Gain Control

The Preamp section controls the gain of the audio signal(s) from the track in which the TimewARP 2600 is running. A rotary knob labeled Gain adjusts the signal level.

If the TimewARP 2600 is running in full stereo confi guration – as a plug-in to a stereo track – the preamp will display two output jacks, one for each stereo channel. Use these signals for any purpose for which you might ordinarily use an

internally-generated signal. You can fi lter them, run them through the Ring Modulator, or use one as an AM or FM program signal. The default input to the Envelope Follower, under these conditions, is taken from the left channel.

4.3 Voltage-Controlled Oscillators (VCO)

These generate three or more of the following basic waveforms. The output amplitude and phase relationships are the same for all oscillators. The oscillator sensitivity under virtual voltage control is 1vV/octave.

For convenience in fi ne-tuning control depth, the three attenuator-governed FM Control inputs at each oscillator provide three different sensitivity ranges. The leftmost slider is full-range; wide open, it passes its signal unchanged. The second slider is 50%; wide open, it passes its signal at half strength. The third slider, wide open, passes its signal at 25% of its original amplitude.

34


4.3.1 VCO 1

VCO 1 generates saw, square, and sine outputs. The sine output is a TimewARP 2600 extension; the original ARP 2600 VCO1 provided just sawtooth and squarewave outputs.

The default signal to the fi rst (unattenuated) FM Control input is from the keyboard. The Audio/LF switch above this input switches the mode of the VCO from Audio (10Hz - 20,000Hz) to LFO Mode (0.03Hz – 30Hz). When the VCO is in LFO Mode, the default connection to the keyboard is removed. This can be overridden in this mode by patching a cable to the Keyboard CV output on the left side of the Keyboard CV output on the left side of the Keyboard CVfront panel.

The default signals to the next three FM Control inputs are from a) the Sample & Hold, b) the ADSR Envelope Generator, and c) VCO2 sine.

4.3.2 VCO 2

VCO 2 generates sine, triangle, sawtooth, and pulse outputs. A pulse-width slide control can adjust the duty cycle from 10% to 90%; at the middle of its travel, the pulse width is 50%, that is, a square wave.

The default signal to the fi rst (unattenuated) FM Control input is from the FM Control input is from the FM Controlkeyboard. The Audio/LF switch above this input switches the mode of the Audio/LF switch above this input switches the mode of the Audio/LF VCOfrom Audio (10Hz - 20,000Hz) to LFO Mode (0.03Hz – 30Hz). When the VCO is in LFO Mode, the default connection to the keyboard is removed. This can be overridden in this mode by patching a cable to the Keyboard CV output on the left side of the Keyboard CV output on the left side of the Keyboard CVfront panel.

The default signals to the next three FM Control inputs are from a) the FM Control inputs are from a) the FM Control Sample & Hold, b) the ADSR Envelope Generator, and c) VCO1 square.

There is a fourth attenuator-governed input, for digital control of the pulse width. The default signal at this PWM input is from the Noise Generator.

CHAPTER 4 -

35


4.3.3 VCO 3

VCO 3 generates sawtooth, pulse, and sine outputs; the pulse width is manually variable. The sine output is a TimewARP 2600 extension; the original ARP 2600 VCO3 provided just sawtooth and pulse outputs.

The default signal to the fi rst (unattenuated) FM Control input is from the keyboard. FM Control input is from the keyboard. FM ControlThe Audio/LF switch above this input switches the mode of the Audio/LF switch above this input switches the mode of the Audio/LF VCO from Audio (10Hz - 20,000Hz) to LFO Mode (0.03Hz – 30Hz). When the VCO is in LFO Mode, the default connection to the keyboard is removed. This can be overridden in this mode by patching a cable to the Keyboard CV output on the left side of the front panel.Keyboard CV output on the left side of the front panel.Keyboard CV

The default signals to the next three FM Control inputs are from a) the Noise Generator,b) the ADSR Envelope Generator, and c) VCO2 sine.

4.4 Voltage Controlled Filter (VCF)

The Voltage Controlled Filter has variable cutoff frequency (Fc) and Voltage Controlled Filter has variable cutoff frequency (Fc) and Voltage Controlled Filterresonance (Q). The response below Fc is fl at down to DC; above Fc the response falls off at 24Db per octave. Fc range is from 10Hz to 10KHz without control voltages; under voltage control, Fc can be driven as far down as 1 Hz and as high as 20KHz.

Fc is controlled manually by a coarse tuning slider (labeled initial fi lter frequency) and a fi ne tune slider. Fc may also be controlled by external voltages; the sensitivity under voltage control is 1.0vV/oct.

The Q, or resonance, of the fi lter circuit is controlled by a single manual slider. As the Q is increased by moving this slider from left to right, the response below Fc is gradually attenuated until a sharp peak remains at the cutoff frequency. (Gain at Fc is always unity.)

At this Q setting, just below the point at which oscillation begins, the fi lter will ring distinctly in response to any sharply defi ned pulse presented to its signal input. In this state it is effectively analogous to

a highly resonant physical system, and may be used for various percussion effects depending on its resonant frequency (identical to Fc) and on the impulse spectrum exciting it.

As the Q is raised still higher, beyond about the halfway point in the slider travel, the fi lter will oscillate. Operating in this state, it generates a pure sine wave. even in the absence of any signal input.

The VCF has fi ve VCF has fi ve VCF Audio signal inputs. They are fed through logarithmic attenuators to a summing point, and then to the VCF itself. The default input signals are from the VCF itself. The default input signals are from the VCF Ring Modulator, VCO-1 Square, VCO-2 Pulse, VCO-3 Sawtooth, and Noise Generator.

The VCF has three frequency Control inputs. The fi rst is normally from the Keyboard pitch-control. The slider that governs this input is a TimewARP 2600 extension; on the original

36


ARP 2600, the keyboard control depth was not adjustable.

The second and third FM Control inputs are governed by linear attenuators; prewired to FM Control inputs are governed by linear attenuators; prewired to FM Controlthese are the ADSR Envelope Generator output and the ADSR Envelope Generator output and the ADSR Envelope Generator VCO-2 Sine output.

Inserting a patch cord at an input jack automatically disconnects the default signal.

4.5 Envelope Generators

The Envelope Generators generate transient, positive-going waveforms, with controllable rise and fall times. They are used primarily with the VCF and VCA, in generating events whose time-varying spectrum and amplitude must be accurately and repeatably controlled. The output from each generator is a positive-going signal whose rise and fall time is set by slide controls on the generator, and whose onset and overall duration is determined by a gate signal.

The maximum value that either envelope can reach is +10vV; thus, unattenuated, either envelope is capable of driving a VCF or VCF or VCF VCA from its minimum initial setting (10Hz for the VCF, -l00Db for the VCA) all the way up to maximum. See 4.1.2 and 4.1.3, specifi cally the data on control input sensitivity. Reread too sections 2.1.6 through 2.1.7.

Gate signals for operating an Envelope Generator may originate with a Envelope Generator may originate with a Envelope Generator Manual Startbutton, the Keyboard Controller, or any +10vV square-wave or pulse signal. The two-position switch just under the lower AR generator selects between the two latter sources. The Manual Start button overrides both of these.

4.5.1 ADSR Envelope Generator

The ADSR Envelope Generator offers variable ADSR Envelope Generator offers variable ADSR Envelope Generator Attack time, initial Decay time, Sustain level, and fi nal Release time. Four vertical sliders control these parameters: note that three of these are time parameters and the fourth – Sustain level – is not.

The generator produces an output only when a gate signal is present at its input. At the onset of a gate signal, the output level rises to +10vV, at a rate set by the Attack slider. When the level reaches +10vV, it immediately begins to fall, at a rate set by the Decayslider, to a level set by the Sustain slider. It remains at this level while the gate signal continues. Finally, when the gate ends, the output falls to zero at a rate set by the fi nal Release slider.

If the gate ends at any time during the ADSR cycle, the generator immediately advances to its fi nal Release phase: the current output level starts falling to zero from whatever value it had at the moment the gate ended.

When the Envelope Generator is controlled from the Envelope Generator is controlled from the Envelope Generator Keyboard Controller, it responds to the trigger signal from the keyboard – supposing that an envelope is already in process – by repeating the fi rst two stages of an envelope and returning again to the Sustain level. For all other control gates presented to the generator input, circuitry internal to the generator itself derives a trigger signal from the leading edge of the gate.

CHAPTER 4 -

37


4.5.2 AR Envelope Generator

The AR Envelope Generator offers variable Attack time and fi nal Release time. Two vertical sliders control these parameters.

The generator produces an output only when a gate signal is present at its input. At the onset of a gate signal, the output level rises to +10vV, at a rate set by the Attack slider. When the level reaches +10vV, it remains at this level while the gate signal continues. Finally, when the gate ends, the output falls to zero at a rate set by the fi nal Releaseslider.

If the gate ends at any time during the AR cycle, the generator immediately advances to its fi nal Release phase: the current output level starts falling to zero from whatever value it had at the moment the gate ended.

The AR Envelope Generator does not require a trigger signal for any phase of its AR Envelope Generator does not require a trigger signal for any phase of its AR Envelope Generatoroperation.

4.6 Voltage Controlled Amplifi er (VCA)

The Voltage Controlled Amplifi er has a maximum gain of unity and a dynamic range Voltage Controlled Amplifi er has a maximum gain of unity and a dynamic range Voltage Controlled Amplifi erof l00Db. With the initial gain control at maximum, and with no control input, theVCA will pass with unchanged amplitude any signal presented to its signal input. On the other hand, with the initial gain control at minimum, no signal will pass through the amplifi er at all unless some positive signal level (the VCA does not respond to negative control signals) is present at one or both of its control inputs.

The fi rst control input has linear sensitivity; the gain of the amplifi er in response to a signal at this input is S/10, i.e. dividing the signal level by 10 will give the gain factor.

The second control input has exponential sensitivity; the gain of the amplifi er in response to a signal at this input will equal 10Db/vV.

There are two audio signal inputs to the VCA. The default connections to these are from the VCF and the Ring Modulator. Inserting a patch cord automatically disconnects the default signal.

The default signal at the linear control input is from the AR Envelope Generator, and at the exponential input is from the ADSR Envelope Generator. Inserting a patch cord automatically disconnects the default signal.

38


4.7 Mix/Pan/Reverb Output Module

The three functions in this module provide fi nal processing of the output signal. That, at least, is what they are intended for; you may actually use them in other roles, for any purpose you please.

If you leave the default connections undisturbed, the module is confi gured as a two-input Mixer, which feeds a Pan Control and a Pan Control and a Pan Control Reverb unit, which are themselves mixed to feed the fi nal left/right system output channels.

When the TimewARP 2600 is confi gured for mono operation, this section omits the Pan control and provides only one output channel.

4.7.1 Mixer

The two inputs to this Mixer carry default connections from the Mixer carry default connections from the Mixer VCF and the VCA; they can of course be overridden with patch cords.

The two jacks just above the sliders and below the Mixer graphic are Mixer graphic are Mixer not inputs; they not inputs; they notare the outputs from the attenuators. This lets you use the two sliders as “fl oating” attenuators, in any situation where you need to set the strength of a signal. (Although if you do this, there’s no way to get any signals into the mixer.)

4.7.2 Pan Control

The Pan Control takes its input from the jack just below the horizontal pan slider. Pan Control takes its input from the jack just below the horizontal pan slider. Pan ControlNormally, this signal comes from the mixer. Centered, the Pan feeds its input signal equally to the left and right channel outputs; moving the slider left or right shifts the signal balance accordingly between the two output channels.

4.7.3 Reverb Unit

The input to this unit is the rightmost jack in the row that runs across the middle of the panel. By default, it carries the Mixer output. The output jack, to its upper right, provides Mixer output. The output jack, to its upper right, provides Mixera 100% wet signal from the Reverb, at a fi xed level. (There are interesting patches in which this signal is subjected to further processing via, say, the Ring Modulator or the Envelope Follower.)

The two sliders adjust the wet-dry mix fed to each output channel.

CHAPTER 4 -

39


4.8 Envelope Follower

The Envelope Follower generates, from any audio-frequency input, a fl uctuating DC output Envelope Follower generates, from any audio-frequency input, a fl uctuating DC output Envelope Followerlevel directly proportional to the average moment-by-moment input signal amplitude. Its sensitivity is such that, with the input attenuator wide open, a 1V P-P square wave will produce a +10vV output. The maximum output is +10vV.

The risetime, or time it takes for the Envelope Follower to respond to any sudden change Envelope Follower to respond to any sudden change Envelope Followerin the amplitude of the signal input, is 10 milliseconds to 50% of fi nal value and 30 milliseconds to 90% of fi nal value.

Like all similar circuitry, the Envelope Follower tends to “ride” on low audio frequencies Envelope Follower tends to “ride” on low audio frequencies Envelope Followeras if they themselves represented changes in signal amplitude; this is not critical, but has been held to a ripple of less than 1% P-P down to l00Hz and less than 10% down to 40Hz.

The primary use of the Envelope Follower is with external instruments. Essentially it extracts, Envelope Follower is with external instruments. Essentially it extracts, Envelope Followerfrom any audio input, a control signal representing the amplitude-envelope of that input: this signal may control the VCF, VCA, or any of the VCO’s. The Envelope Follower output is an Envelope Follower output is an Envelope Followerenvelope and can be used in the same fashion as the output from either of the envelope generators.

The default input to the Envelope Follower is from the preamplifi er. When the TimewARP Envelope Follower is from the preamplifi er. When the TimewARP Envelope Follower2600 is confi gured for stereo input, it is the fi rst (left) channel preamplifi er output.

40


4.9 Ring Modulator

The Ring Modulator is essentially a multiplier; from its two inputs A and B it produces the Ring Modulator is essentially a multiplier; from its two inputs A and B it produces the Ring Modulatoroutput function A x B / 5. The kind of transformation this effects on input signals depends to a large extent on what they are and on whether the modulator is AC or DC coupled to them. This is selected by the Audio/DC switch at the bottom of the modulator.Audio/DC switch at the bottom of the modulator.Audio/DC

When the inputs are AC coupled (Audio When the inputs are AC coupled (Audio When the inputs are AC coupled ( position of the switch), any DC component present in them is canceled before they are fed to the modulator. Thus a sawtooth that starts from zero and goes to+l0vV will instead start at -5vV and move to +5vV so that its overall positive and negative deviation cancels to zero. Under these conditions the modulator will generate from any two periodic signals an output signal consisting of the sum and difference frequencies that can be generated from the frequencies of the two inputs. The input frequencies themselves will be suppressed.

If both signals are audio-frequency, a large variety of harmonic and inharmonic timbres can be produced from the modulator, depending on the ratio of the input frequencies and on their own harmonic content. If A is a sine wave and we represent its frequency by Fa, and B is a complex waveform of frequency Fb with overtones 2Fb, 3Fb, 4Fb, etc., then the output of the modulator will be a complex waveform with frequency components Fb + Fa, Fb - Fa, 2Fb±Fa, 3Fb ± Fa, 4Fb ± Fa, etc. A moment’s experimentation with the prewired sawtooth and sine inputs to the modulator will demonstrate the complexity of the timbres that can be generated by this simple means.

If, still with AC coupling, one input is subsonic and the other at some audio frequency, there will be an output from the modulator only when the value of the subsonic input is changing, and the output will be roughly proportional to the rate of change. If, for example, the subsonic input is a square wave, the modulator output will be a series of short, decaying tonebursts – one at each rise or fall in the input signal.

When the inputs are DC coupled, any DC component in either one of the inputs will pass into the modulator and affect the modulating process. The effect when both inputs are at audio frequency is to allow into the output waveform some of the input frequencies in addition to the sum and difference frequencies. The effect when one of the inputs is subsonic is that the modulator operates as a voltage-controlled amplifi er: the output amplitude will be in direct proportion to the instantaneous amplitude of the low-frequency input and will vary as its absolute value varies. Also, the output phase will reverse when the low-frequency input signal changes from positive to negative or vice versa.

The AC-coupling time constants are 235 msec, for the left input and 90 msec, for the right input.

CHAPTER 4 -

41


4.10 Noise Generator (NG)

The Noise Generator has two manual controls: one for spectral balance (“color”) and one Noise Generator has two manual controls: one for spectral balance (“color”) and one Noise Generatorfor output level.

The spectral balance is continuously variable from white to red (low-frequency noise output). In the latter case the output falls off at the rate of 6Db/Octave; the pink noise position approximates a -3Db/Octave slope.

The level control, at minimum, cuts off the output signal completely. At maximum, the output is clipped at 20vV P-P to produce binary, or two-valued, noise. Clipping begins with the level control approximately half open. input.

4.11 Voltage Processors (VP)

The Voltage Processors are simple utility functions for mixing, inverting, and shaping signals.

4.11.1 VP #1

VP #1 has four signal inputs and one output. Two of the inputs have attenuators. The output signal is the inverted sum of all four inputs.

The attenuator-governed inputs carry default connections from +10vV and from the keyboard pitch-control. Opening the +10vV slider thus produces up to -10vV at the processor output.

4.11.2 VP #2

VP #2 has two signal inputs and one output. One of the inputs has an attenuator. The output signal is the inverted sum of the two inputs.

The attenuator-governed input carries a default connection from -10vV, so that opening the slider produces up to +10vV at the output.

4.11.3 The Lag Processor

The Lag Processor is a low-pass fi lter for processing control signals. The slider adjusts Lag Processor is a low-pass fi lter for processing control signals. The slider adjusts Lag Processorits cutoff frequency. The corresponding rise-time ranges from 0.5ms with the slider at minimum, to 500msec – about half a second – with the slider at maximum.

The Lag Processor can be used to process audio signals, as a -6Db/octave manual fi lter Lag Processor can be used to process audio signals, as a -6Db/octave manual fi lter Lag Processorwith a maximum Fc of approximately 1KHz.

42


4.12 The Sample & Hold Module (S/H)

The Sample & Hold Module produces stepped output signal levels, by sampling the instantaneous value of any signal at its input. The stepped levels produced in this manner are useful for controlling oscillator and fi lter frequencies and – occasionally – VCA gain.

The S/H circuit has a signal input (the waveform to be sampled), a signal input (the waveform to be sampled), a signal trigger input, and an output giving the result of the sampling operation. The trigger input is defaulted from the internal clock, but any square or pulse wave, or the keyboard

gate or trigger signals, will work.

Upon being triggered, the S/H sets its output level to the same value as the input signal at that instant. After the trigger, the output signal will hold that level until the next trigger pulse.

Any signal whatsoever may be sampled. The default input is from the Noise Generator,so that the step sequence is random. The accompanying diagrams show how, when the signal being sampled is random noise, the output voltages are correspondingly unpredictable. An infi nite variety of cyclical output patterns may be obtained, on the other hand, by sampling a periodic waveform. Different ratios of the sampling frequencies to the frequency of the waveform being sampled create different melodic patterns (if the output level is controlling a VCO).

The level control attenuates the input signal before it is fed to the S/H circuit. The rate control actually belongs to the internal clock; when that is disconnected from the S/Hcircuit, the rate control has no effect on the operation of the S/H circuit.

4.13 The Internal Clock / Electronic Switch

The Internal Clock is a manually controlled low-frequency square-wave oscillator. It is the Internal Clock is a manually controlled low-frequency square-wave oscillator. It is the Internal Clockdefault trigger source for the S/H device. It is also hardwired as the clock source for the S/H device. It is also hardwired as the clock source for the S/HElectronic Switch.

Under MIDI control, the Internal Clock may be synchronized to incoming MIDI Beat Clocks; see section 4.1.11.4.

The Electronic Switch has two connections on one side and one on the other, as indicated by the panel graphics. For clarity, let’s call these three jacks A-1, A-2, and B. The switch alternates between connecting A-1 to B, and A-2 to B. It doesn’t matter which side is the signal source and which is the destination; the switch works the same regardless.

The switching rate is governed by the Internal Clock. This is a permanent feature of the switch.

CHAPTER 4 -

43


4.14 The Virtual Keyboard

This has fi ve octaves, 60 keys. The control panel, at the left, is modeled on but not identical to the original ARP 3620 keyboard.

4.14.1 Low Frequency Oscillator (LFO) Section

The keyboard unit has its own LFO section, independent of any of the standard VCO’s. It can be used in two ways: for vibrato, or for automatically repeated keyboard gates (as, for example, in imitating the repeated notes of a mandolin). Three sliders govern the Speed, Delay, and Depth of the LFO.

Under MIDI control, the keyboard LFO may be synchronized to incoming MIDIBeat Clock signals; see section 4.1.11.4.

4.14.2 Dual-Pitch Control Output

Like the original ARP 2600, the TimewARP 2600 virtual keyboard can generate a second pitch-control signal when two keys are depressed. This signal is available at the two jacks labeled Upper Voice at the lower left of the keyboard module.

To use one of these, simply patch it to an oscillator. That oscillator will now track the uppermost key depressed rather than – as with the standard keyboard control signal – the lower key.

4.14.3 Gate and Trigger Control

Two switches in the upper right quadrant of the keyboard module govern the logic of the keyboard gate and trigger signals.

When the Trigger Mode switch is set to Off, the keyboard generates a continuous gate signal as long as any key is depressed, and generates a trigger signal only on the transition from no key depressed to any key depressed. In this operating mode, you have complete performing control over the production of trigger signals; to avoid them, play legato, and to generate them, play non-legato. This is the baseline logic of the original ARP 2600 keyboard.

With this switch set to On, the keyboard will generate a trigger on every new keypress, regardless of your performing habits. The gate logic is not affected.

The three-position switch labeled Auto Repeat isAuto Repeat isAuto Repeat Off in its center position. This is the Off in its center position. This is the Offdefault.

In its lower position, the keyboard gate and trigger are taken from the local LFO. Actual key depressions no longer play a role in gating. In its upper position, the LFO and the keypress are ANDed together; when you press a key, there is a series of pulses from the LFO, and when you release the key, the series stops. This is the mandolin effect we mentioned above.

The keyboard Gate and Trigger signals are available on the main panel, from two jacks in Trigger signals are available on the main panel, from two jacks in Triggerthe Envelope Generator section.Envelope Generator section.Envelope Generator

Please see the Patchman.pdf fi le found on the install disc or at www.wayoutware.com for suggestions. Patchman.pdf fi le found on the install disc or at www.wayoutware.com for suggestions. Patchman.pdfIf you are new to audio synthesis, consult the tutorial patch collection.

Patching the TimewARP 2600 5

6.1 Table of Alternate keyboard tunings

Tuning Presets, compiled by Robert Rich

6.1.1 12 Tone Equal Temperament (non-erasable)

The default Western tuning, based on the twelfth root of two. Good fourths and fi fths, horrible thirds and sixths.

6.1.2 Harmonic Series

MIDI notes 36-95 refl ect harmonics 2 through 60 based on the fundamental of A = 27.5 Hz. The low C on a standard 5 octave keyboard acts as the root note (55Hz), and the harmonics play upwards from there. The remaining keys above and below the 5 octave range are fi lled with the same intervals as Carlos’ Harmonic 12 Tone that follows.

6.1.3 Carlos Harmonic Twelve Tone

Wendy Carlos’ twelve note scale based on octave-repeating harmonics.

A = 1/1 (440 Hz).

1/1 17/16 9/8 19/16 5/4 21/16 11/8 3/2 13/8 27/16 7/4 15/8

6.1.4 Meantone Temperament

An early tempered tuning, with better thirds than 12ET. Sounds best in the key of C. Use this to add an authentic touch to performances of early Baroque music. C = 1/1 (260 Hz)

6.1.5 1/4 Tone Equal Temperament

24 notes per octave, equally spaced 24root2 intervals. Mexican composer Julian Carillo used this for custom-built pianos in the early 20th century.

6.1.6 19 Tone Equal Temperament

19 notes per octave (19root2) offering better thirds than 12 ET, a better overall compromise if you can fi gure out the keyboard patterns.

6Appendices

48

CHAPTER 6 - Appendices CHAPTER 6 -

6.1.7 31 Tone Equal Temperament

Many people consider 31root2 to offer the best compromise towards just intonation in an equal temperament, but it can get very tricky to keep track of the intervals.

6.1.8 Pythagorean C

One of the earliest tuning systems known from history, the Pythagorean scale is constructed from an upward series of pure fi fths (3/2) transposed down into a single octave. The tuning works well for monophonic melodies against fi fth drones, but has a very narrow palate of good chords to choose from. .

C = 1/1 (261.625 Hz)

1/1 256/243 9/8 32/27 81/64 4/3 729/512 3/2 128/81 27/16

16/9 243/128

6.1.9 Just Intonation in A with 7-limit Tritone at D#

A rather vanilla 5-limit small interval JI, except for a single 7/5 tritone at D#, which offers some nice possibilities for rotating around bluesy sevenths.

A = 1/1 (440 Hz)

1/1 16/15 9/8 6/5 5/4 7/5 3/2 8/5 5/3 9/5 15/8

6.1.10 3-5 Lattice in A

A pure 3 and 5-limit tuning which resolves to very symmetrical derived relationships between notes. A = 1/1 (440 Hz)

1/1 16/15 10/9 6/5 5/4 4/3 64/45 3/2 8/5 5/3 16/9 15/8

6.1.11 3-7 Lattice in A

A pure 3 and 7-limit tuning which resolves to very symmetrical derived relationships between notes. Some of the intervals are very close together, offering several choices for the same nominal chords.

A= 1/1 (440 Hz)

1/1 9/8 8/7 7/6 9/7 21/16 4/3 3/2 32/21 12/7 7/4 63/32

6.1.12 Other Music 7-Limit Black Keys in C

Created by the group Other Music for their homemade gamelan, this offers a wide range of interesting chords and modes.

C= 1/1 (261.625 Hz)

1/1 15/14 9/8 7/6 5/4 4/3 7/5 3/2 14/9 5/3 7/4 15/8

CHAPTER 6 -

49

CHAPTER 6 - Appendices

6.1.13 Dan Schmidt Pelog/Slendro

Created for the Berkeley Gamelan group, this tuning fi ts an Indonesian-style heptatonic Pelog on the white keys and pentatonic Slendro on the black keys, with B and Bb acting as 1/1 for their respective modes. Note that some of the notes will have the same frequency. By tuning the 1/1 to 60 Hz, Dan found a creative way to incorporate the inevitable line hum into his scale.

Bb, B = 1/1 (60 Hz)

1/1 1/1 9/8 7/6 5/4 4/3 11/8 3/2 3/2 7/4 7/4 15/8

6.1.14 Yamaha Just Major C

When Yamaha decided to put preset microtunings into their FM synth product line, they selected this and the following tuning as representative just intonations. As such, they became the de-facto introduction to JI for many people. Just Major gives preferential treatment to major thirds on the sharps, and a good fourth relative to the second.

C= 1/1 (261.625)

1/1 16/15 9/8 6/5 5/4 4/3 45/32 3/2 8/5 5/3 16/9 15/8

6.1.15 Yamaha Just Minor C

Similar to Yamaha=92s preset Just Major, the Just Minor gives preferential treatment to minor thirds on the sharps, and has a good fi fth relative to the second.

C= 1/1 (261.625)

1/1 25/24 10/9 6/5 5/4 4/3 45/32 3/2 8/5 5/3 16/9 15/8

6.1.16 Harry Partch 11-limit 43 Note Just Intonation

One of the pioneers of modern microtonal composition, Partch built a unique orchestra with this tuning during the fi rst half of the 20th century, to perform his own compositions. The large number of intervals in this very dense scale offers a full vocabulary of expressive chords and complex key changes. The narrow spacing also allows fi xed-pitched instruments like marimbas and organs to perform glissando-like passages.

G = 1/1 (392 Hz, MIDI note 67)

1/1 81/80 33/32 21/20 16/15 12/11 11/10 10/9 9/8 8/7 7/6

32/27 6/5 11/9 5/4 14/11 9/7 21/16 4/3 27/20 11/8 7/5

10/7 16/11 40/27 3/2 32/21 14/9 11/7 8/5 18/11 5/3 27/16 12/7

7/4 16/9 9/5 20/11 11/6 15/8 40/21 64/33 160/81

Index 7

Documents

OWNER’S MANUALcrosstalk.arizona.edu/Manuals/Timewarp.pdfEdgar Winter, Paul Bley, Roger Powell, Jean-Michel Jarre, Mike Oldfi eld, Herbie Hancock, and many, many others. Its modular